u
This commit is contained in:
sebseb7
247
generate-category-descriptions.js
Normal file
247
generate-category-descriptions.js
Normal file
@@ -0,0 +1,247 @@
|
||||
#!/usr/bin/env node
|
||||
|
||||
import fs from 'fs';
|
||||
import path from 'path';
|
||||
import OpenAI from 'openai';
|
||||
|
||||
// Configuration
|
||||
const OPENAI_API_KEY = process.env.OPENAI_API_KEY;
|
||||
const DIST_DIR = './dist';
|
||||
const OUTPUT_CSV = './category-descriptions.csv';
|
||||
|
||||
// Model configuration
|
||||
const MODEL = 'gpt-5.1';
|
||||
|
||||
// Initialize OpenAI client
|
||||
const openai = new OpenAI({
|
||||
apiKey: OPENAI_API_KEY,
|
||||
});
|
||||
|
||||
// System prompt for generating SEO descriptions
|
||||
const SEO_DESCRIPTION_PROMPT = `You are given a list of products from a specific category. Create a SEO-friendly description for that category that would be suitable for a product catalog page.
|
||||
|
||||
Requirements:
|
||||
- Write in German
|
||||
- Make it SEO-optimized with relevant keywords
|
||||
|
||||
The product list format is:
|
||||
First line: categoryName,categoryId
|
||||
Subsequent lines: articleNumber,price,productName,shortDescription
|
||||
|
||||
Generate a compelling category description based on this product data.`;
|
||||
|
||||
// Function to find all *-list.txt files in dist directory
|
||||
function findListFiles() {
|
||||
try {
|
||||
const files = fs.readdirSync(DIST_DIR);
|
||||
return files.filter(file => file.endsWith('-list.txt'));
|
||||
} catch (error) {
|
||||
console.error('Error reading dist directory:', error.message);
|
||||
return [];
|
||||
}
|
||||
}
|
||||
|
||||
// Function to read a list file and extract category info
|
||||
function readListFile(filePath) {
|
||||
try {
|
||||
const content = fs.readFileSync(filePath, 'utf8');
|
||||
const lines = content.trim().split('\n');
|
||||
|
||||
if (lines.length < 1) {
|
||||
throw new Error('File is empty');
|
||||
}
|
||||
|
||||
// Parse first line: categoryName,categoryId
|
||||
const firstLine = lines[0];
|
||||
const [categoryName, categoryId] = firstLine.split(',');
|
||||
|
||||
if (!categoryName || !categoryId) {
|
||||
throw new Error('Invalid first line format');
|
||||
}
|
||||
|
||||
return {
|
||||
categoryName: categoryName.replace(/^"|"$/g, ''), // Remove quotes if present
|
||||
categoryId: categoryId.replace(/^"|"$/g, ''),
|
||||
content: content
|
||||
};
|
||||
} catch (error) {
|
||||
console.error(`Error reading ${filePath}:`, error.message);
|
||||
return null;
|
||||
}
|
||||
}
|
||||
|
||||
// Function to generate SEO description using OpenAI
|
||||
async function generateSEODescription(productListContent, categoryName, categoryId) {
|
||||
try {
|
||||
console.log(`🔄 Generating SEO description for category: ${categoryName} (ID: ${categoryId})`);
|
||||
|
||||
const response = await openai.responses.create({
|
||||
model: "gpt-5.1",
|
||||
input: [
|
||||
{
|
||||
"role": "developer",
|
||||
"content": [
|
||||
{
|
||||
"type": "input_text",
|
||||
"text": SEO_DESCRIPTION_PROMPT
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"role": "user",
|
||||
"content": [
|
||||
{
|
||||
"type": "input_text",
|
||||
"text": productListContent
|
||||
}
|
||||
]
|
||||
}
|
||||
],
|
||||
text: {
|
||||
"format": {
|
||||
"type": "json_schema",
|
||||
"name": "descriptions",
|
||||
"strict": true,
|
||||
"schema": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"seo_description": {
|
||||
"type": "string",
|
||||
"description": "A concise description intended for SEO purposes. 155 characters"
|
||||
},
|
||||
"long_description": {
|
||||
"type": "string",
|
||||
"description": "A comprehensive description, 2-5 Sentences"
|
||||
}
|
||||
},
|
||||
"required": [
|
||||
"seo_description",
|
||||
"long_description"
|
||||
],
|
||||
"additionalProperties": false
|
||||
}
|
||||
},
|
||||
"verbosity": "medium"
|
||||
},
|
||||
reasoning: {
|
||||
"effort": "none",
|
||||
"summary": "auto"
|
||||
},
|
||||
tools: [],
|
||||
store: false,
|
||||
include: [
|
||||
"reasoning.encrypted_content",
|
||||
"web_search_call.action.sources"
|
||||
]
|
||||
});
|
||||
|
||||
const description = response.output_text;
|
||||
console.log(`✅ Generated description for ${categoryName}`);
|
||||
return description;
|
||||
|
||||
} catch (error) {
|
||||
console.error(`❌ Error generating description for ${categoryName}:`, error.message);
|
||||
return `Error generating description: ${error.message}`;
|
||||
}
|
||||
}
|
||||
|
||||
// Function to write CSV file
|
||||
function writeCSV(results) {
|
||||
try {
|
||||
const csvHeader = 'categoryId,listFileName,seoDescription\n';
|
||||
const csvRows = results.map(result =>
|
||||
`"${result.categoryId}","${result.listFileName}","${result.description.replace(/"/g, '""')}"`
|
||||
).join('\n');
|
||||
|
||||
const csvContent = csvHeader + csvRows;
|
||||
fs.writeFileSync(OUTPUT_CSV, csvContent, 'utf8');
|
||||
console.log(`✅ CSV file written: ${OUTPUT_CSV}`);
|
||||
console.log(`📊 Processed ${results.length} categories`);
|
||||
|
||||
} catch (error) {
|
||||
console.error('Error writing CSV file:', error.message);
|
||||
}
|
||||
}
|
||||
|
||||
// Main execution function
|
||||
async function main() {
|
||||
console.log('🚀 Starting category description generation...');
|
||||
|
||||
// Check if OpenAI API key is set
|
||||
if (!OPENAI_API_KEY) {
|
||||
console.error('❌ OPENAI_API_KEY environment variable is not set');
|
||||
console.log('Please set your OpenAI API key: export OPENAI_API_KEY="your-api-key-here"');
|
||||
process.exit(1);
|
||||
}
|
||||
|
||||
// Check if dist directory exists
|
||||
if (!fs.existsSync(DIST_DIR)) {
|
||||
console.error(`❌ Dist directory not found: ${DIST_DIR}`);
|
||||
process.exit(1);
|
||||
}
|
||||
|
||||
// Find all list files
|
||||
const listFiles = findListFiles();
|
||||
if (listFiles.length === 0) {
|
||||
console.log('⚠️ No *-list.txt files found in dist directory');
|
||||
console.log('💡 Make sure to run the prerender script first to generate the list files');
|
||||
process.exit(1);
|
||||
}
|
||||
|
||||
console.log(`📂 Found ${listFiles.length} list files to process`);
|
||||
|
||||
const results = [];
|
||||
|
||||
// Process each list file
|
||||
for (const listFile of listFiles) {
|
||||
const filePath = path.join(DIST_DIR, listFile);
|
||||
|
||||
// Read and parse the file
|
||||
const fileData = readListFile(filePath);
|
||||
if (!fileData) {
|
||||
console.log(`⚠️ Skipping ${listFile} due to read error`);
|
||||
continue;
|
||||
}
|
||||
|
||||
// Generate SEO description
|
||||
const description = await generateSEODescription(
|
||||
fileData.content,
|
||||
fileData.categoryName,
|
||||
fileData.categoryId
|
||||
);
|
||||
|
||||
// Store result
|
||||
results.push({
|
||||
categoryId: fileData.categoryId,
|
||||
listFileName: listFile,
|
||||
description: description
|
||||
});
|
||||
|
||||
// Add delay to avoid rate limiting
|
||||
await new Promise(resolve => setTimeout(resolve, 1000));
|
||||
}
|
||||
|
||||
// Write CSV output
|
||||
if (results.length > 0) {
|
||||
writeCSV(results);
|
||||
console.log('🎉 Category description generation completed successfully!');
|
||||
} else {
|
||||
console.error('❌ No results to write - all files failed processing');
|
||||
process.exit(1);
|
||||
}
|
||||
}
|
||||
|
||||
// Run the script
|
||||
if (import.meta.url === `file://${process.argv[1]}`) {
|
||||
main().catch(error => {
|
||||
console.error('❌ Script failed:', error.message);
|
||||
process.exit(1);
|
||||
});
|
||||
}
|
||||
|
||||
export {
|
||||
findListFiles,
|
||||
readListFile,
|
||||
generateSEODescription,
|
||||
writeCSV
|
||||
};
|
||||
Reference in New Issue
Block a user